A knee brace added a small, fading benefit on top of exercise advice in 466 adults

Knee braces for osteoarthritis occupy an odd position. They are sold widely, recommended inconsistently, and have rarely been tested at scale against the thing patients are actually offered first, which is advice and exercise. A trial published in The BMJ on 26 January addresses that directly, and finds a benefit that is real, small, and does not last well [s1].

The comparison

PROP OA was a multicentre, parallel group, superiority trial with a blinded statistician, recruiting through general practices and the community in Cheshire, Manchester, North Tyneside and Staffordshire in England [s1]. Enrolment ran from 25 November 2019 to 16 September 2022, and 466 adults aged 45 or older with symptoms of knee osteoarthritis were randomised 1:1 [s1]. Mean age was 64 (SD 9) and 46% of participants were women [s1].

Both groups received the same baseline: a single in-person consultation with a trained physiotherapist covering advice, written information and exercise instruction [s1]. The intervention group was additionally fitted with a knee brace selected to match where their osteoarthritis predominantly sat — patellofemoral, tibiofemoral unloading, or neutral stabilising — and offered a two-week follow-up consultation [s1]. Brief motivational interviewing and targeted text reminders were used to support brace adherence [s1].

That adherence component is worth flagging. What was tested is not "a brace"; it is a brace plus a fitting consultation plus a behavioural support package.

The numbers

The primary outcome was the composite Knee Osteoarthritis Outcomes Score (KOOS-5), scored 0 to 100, at six months [s1]. Follow-up with analysable data was 401 participants (86%) at three months, 394 (85%) at six, and 370 (79%) at twelve [s1].

At six months, the brace group improved more: an adjusted mean difference of 3.39 points (95% CI 0.96 to 5.82), an effect size of 0.24 [s1]. The confidence interval excludes zero, so the result is statistically robust, but 3.39 points on a 100-point scale is a small difference and the authors describe it as such [s1].

Pain did better than the composite. On the KOOS pain subscale, also 0 to 100, the adjusted mean difference at six months was 6.13 points (95% CI 3.36 to 8.91), effect size 0.39 [s1]. This was the largest effect observed [s1].

The direction of travel over time is the finding most likely to matter to someone deciding whether to try one. Secondary outcomes showed benefits of the brace group over the advice-only group that diminished over time [s1]. Adverse events were minor and expected [s1].

The authors' conclusion is that adding compartment-specific bracing with an adherence intervention resulted in small improvements, and that this safe intervention offers a potential treatment option [s1].

Why this lands in a gap

A systematic review published in Bone & Joint Open in November 2025 mapped what international guidelines actually say about non-surgical, non-pharmacological management of knee osteoarthritis [s2]. Searching CINAHL, MEDLINE, Science Direct and the Guideline International Network website under PRISMA guidance, the authors identified 13 current clinical practice guidelines from national and international medical, surgical and professional associations, and rated the quality of guideline construction as high [s2].

On the core of management, the guidelines agree: exercise, self-management advice and education, and weight management [s2]. Beyond that core, they diverge — and bracing and orthotics are named among the interventions where notable variation exists, alongside manual therapy and acupuncture [s2]. The reviewers attribute the divergence not to different evidence but to different interpretations of generally low levels of evidence, and note this variation may be confusing for patients and clinicians alike [s2]. They also found the guidelines lacked detail on how the core interventions should actually be delivered [s2].

PROP OA is the sort of trial that could reduce that divergence, because it tests bracing on top of the core that every guideline already agrees on rather than against nothing.

What it does not establish

The trial compared brace-plus-package against a single physiotherapy consultation. It cannot separate the contribution of the brace from the fitting appointment, the follow-up consultation, the motivational interviewing or the text reminders. It does not compare brace types against each other. Twelve months is the limit of the reported follow-up, and by then the benefits were diminishing [s1]. This is one trial in one country's health system, and the population was recruited from primary care and the community rather than from surgical waiting lists.

Nothing here tells any individual reader whether a brace is appropriate for their knee, and the compartment-specific fitting that the trial used is not something a person can do from a product listing.

What to watch

Whether guideline groups revisit their bracing recommendations now that a trial exists testing bracing as an addition to standard advice rather than as an alternative to it, and whether longer follow-up shows the effect disappearing entirely or stabilising at a low level.

Sources

1. [s1] Provision of knee bracing for knee osteoarthritis (PROP OA): multicentre, parallel group, superiority, statistician blinded, randomised controlled trial. BMJ, 26 January 2026. https://doi.org/10.1136/bmj-2025-086005
2. [s2] Consistency of advice for the conservative management of knee osteoarthritis across international clinical practice guidelines. Bone & Joint Open, 4 November 2025. https://doi.org/10.1302/2633-1462.611.BJO-2024-0153.R1